拓扑与资源感知的虚拟网络功能迁移方法

doi:10.3778/j.issn.1673-9418.2008017

摘要/Abstract

摘要：

针对网络功能虚拟化环境中网络出现负载失衡的问题，提出一种拓扑与资源感知的虚拟网络功能迁移方法（TRA-VNFM）。首先，根据底层网络的计算、存储和转发资源占用情况，设置两级动态阈值对物理节点的过载程度进行分类，同时制定相应的迁移判定条件并计算出待迁移目的节点集。其中高过载节点优先实施迁移且有更低的迁移成功条件。其次，针对过载节点上部署的虚拟网络功能，利用资源感知算法对其迁移权重进行设定，占用过载资源越多的虚拟网络功能迁移权重越大，结合迁移权重和资源需求选择出待迁移虚拟网络功能。最后，采用极值交互的拓扑感知算法综合考虑各类资源占用情况、处理时延以及拓扑属性，对待迁移目的节点集中的节点进行评价，将评价最高的节点作为迁移目的节点。仿真实验表明，与以往虚拟网络功能迁移方法相比，该方法不仅降低了迁移时间，还在服务功能链的平均时延、网络的收益开销比与负载均衡程度方面有较好的性能。

关键词: 网络功能虚拟化（NFV）, 虚拟网络功能（VNF）迁移, 负载均衡, 极值交互, 拓扑与资源感知

Abstract:

In the network function virtualization environment, aiming at the load imbalance of network, a topology and resource-aware virtual network function migration method (TRA-VNFM) is proposed. Firstly, according to the computing, storage and forwarding resource occupancy of the underlying network, the two-level dynamic threshold is set to classify the overload degree of physical nodes, and at the same time, the corresponding migration judgment conditions are formulated and the set of destination nodes to be migrated is calculated. Among them, high-overload nodes have priority to implement migration and have lower migration success conditions. Secondly, for the virtual network function deployed on the overload node, the resource-aware algorithm is used to set its migration weight.The more overload resources the virtual network function occupies, the larger its migration weight, combining the migration weight and resource demand to select the virtual network function to be migrated. Finally, the topology-aware algorithm of extreme value interaction is used to evaluate the nodes in the migration destination node set, taking the node with the highest evaluation as the migration destination node by considering all kinds of resource occupancy, processing delay and topological properties. Simulation experiments show that compared with the previous virtual network function migration methods, this method not only reduces the migration time, but also has better performance in the average delay of the service function chain, the revenue to expense ratio of network and the degree of load balancing.

Key words: network function virtualization (NFV), virtual network function (VNF) migration, load balancing, extreme value interaction, topology and resource awareness

阳勇，孟相如，康巧燕，韩晓阳. 拓扑与资源感知的虚拟网络功能迁移方法[J]. 计算机科学与探索, 2021, 15(11): 2161-2170.

YANG Yong, MENG Xiangru, KANG Qiaoyan, HAN Xiaoyang. Virtual Network Function Migration Method Based on Topology and Resource Awareness[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(11): 2161-2170.

参考文献

[1] HEIDEKER A, KAMIENSKI C A. ElasticNFV: an elasticity manager for NFV using SDN[J]. IEEE Latin America Trans-actions, 2019, 17(1): 167-173.
[2] LINGUAGLOSSA L, LANGE S, PONTARELLI S, et al. Survey of performance acceleration techniques for network function virtualization[J]. Proceedings of the IEEE, 2019, 107(4): 746-764.
[3] KARIMZADEH-FARSHBAFAN M, SHAH-MANSOURI V, NIYATO D. A dynamic reliability-aware service placement for network function virtualization (NFV)[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(2): 318-333.
[4] HAWILO H, JAMMAL M, SHAMI A. Network function virtualization-aware orchestrator for service function chaining placement in the cloud[J]. IEEE Journal on Selected Areas in Communications, 2019, 37(3): 643-655.
[5] BUYAKAR T V K, AGARWAL H, TAMMA B R, et al. Proto-typing and load balancing the service based architecture of 5G core using NFV[C]//Proceedings of the 5th IEEE Conference on Network Softwarization, Paris, Jun 24-28, 2019. Piscataway: IEEE, 2019: 228-232.
[6] ATTIAH K M, BANAWAN K, GABER A, et al. Load bal-ancing in cellular networks: a reinforcement learning approach[C]//Proceedings of the IEEE 17th Annual Consumer Com-munications & Networking Conference, Las Vegas, Jan 10-13, 2020. Piscataway: IEEE, 2020: 1-6.
[7] PEI J N, HONG P L, PAN M, et al. Optimal VNF placement via deep reinforcement learning in SDN/NFV-enabled net-works[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(2): 263-278.
[8] TANG L, ZHAO P P, ZHAO G F, et al. Dynamic deployment algorithm for service function chaining with QoS guarantee[J]. Journal of Beijing University of Posts and Telecommu-nications, 2018, 41(6): 90-96.
唐伦, 赵培培, 赵国繁, 等. 基于QoS保障的服务功能链动态部署算法[J]. 北京邮电大学学报, 2018, 41(6): 90-96.
[9] YOU C Q, LI L M. Efficient load balancing for the VNF deployment with placement constraints[C]//Proceedings of the 2019 IEEE International Conference on Communications, Shanghai, May 20-24, 2019. Piscataway: IEEE, 2019: 1-6.
[10] TANG L, YANG H, MA R L, et al. Multi-priority based joint optimization algorithm of virtual network function migration cost and network energy consumption[J]. Journal of Elec-tronics & Information Technology, 2019, 41(9): 2079-2086.
唐伦, 杨恒, 马润琳, 等. 基于5G接入网络的多优先级虚拟网络功能迁移开销与网络能耗联合优化算法[J]. 电子与信息学报, 2019, 41(9): 2079-2086.
[11] WANG Y Z, ZHU X R, QIU X H. A quick adaptive migration algorithm for virtual network function[C]//Proceedings of the 10th EAI International Conference on Wireless and Sate-llite Systems, Harbin, Jan 12-13, 2019. Berlin, Heidelberg: Springer, 2019: 333-347.
[12] ZHOU X, YI B, WANG X, et al. Approach for minimizing network effect of VNF migration[J]. IET Communications, 2018, 12(20): 2574-2581.
[13] EJAZ S, IQBAL Z, SHAH P A, et al. Traffic load balancing using software defined networking (SDN) controller as virtua-lized network function[J]. IEEE Access, 2019, 7: 46646-46658.
[14] ZHU X R, WANG Y Z. Research on virtual network function (VNF) migration[J]. Journal of Nanjing University of Posts and Telecommunications (Natural Science Edition), 2018, 38(1): 45-53.
朱晓荣, 王一忠. 虚拟网络功能迁移研究综述[J]. 南京邮电大学学报(自然科学版), 2018, 38(1): 45-53.
[15] WANG Y W, LIU M, MA C, et al. High performance load balancing mechanism for network function virtualization[J]. Journal of Computer Research and Development, 2018, 55(4): 689-703.
王煜炜, 刘敏, 马诚, 等. 面向网络功能虚拟化的高性能负载均衡机制[J]. 计算机研究与发展, 2018, 55(4): 689-703.